I build products. AI multiplies, it doesn't replace.
More than ten years building applications, five of them with AI. I work across the whole cycle: scoping, UX/UI, frontend, backend, data, authentication, APIs, workflows, AI and deployment.
A hybrid profile
Product+ UX/UI+ Development+ Data+ Automation+ AI= a product that runs.
Usually, that's six people and as many hand-offs. I bring them together, so nothing gets lost between two disciplines.
What I master. What I'm experimenting with. What I'm learning.
Telling the three apart is the least I can do when you trust me with a project.
I master
- Supabase: database, authentication, storage, access rules
- n8n, REST APIs and webhooks
- Claude Code, Codex and Lovable to develop fast
- UX/UI and product prototyping
- Deployment on Vercel, Hostinger, OVH or a VPS
- GitHub and continuous delivery
I'm experimenting with
- AI agents that chain several actions
- Evaluating the quality of an AI's answers
- AI-assisted visual creation with Higgsfield
- WebGL and shaders, as on this website
I'm learning
- Very high-load architectures
- Native mobile development
- Advanced security certifications
When your project depends on them, I tell you before we start.
Small teams, augmented by AI.
One person who thinks in product terms and knows how to use AI can now deliver what took a small team not so long ago.
What can't be automated: understanding the problem, making choices, standing behind the quality of what you deliver. That is where I spend my time.
How I work.
- Precise
- I show what is built, and why.
- Clear
- Technology explained simply, without dumbing it down.
- Results-driven
- The business problem comes before the tool.
- Visual
- Demos, diagrams, prototypes: you see before you approve.
- Credible
- No promise I can't keep.
Also at the head of two agencies.
That is where I learned what a manual process costs, and what a system that runs is worth.